AI-generated conjecture · below the evidence/publication boundary

Heredity is scale's default; the throne was the distraction

Status: Falsified

Status is derived only from the shepherd-authored triage/prediction data above -- community submissions and claims are a separate overlay and can never change it (see the participation panel below).

This is a proposed connection between two domains, generated by a language model. It is not an article and not evidence: it sits below the evidence/publication boundary. A quantitative prediction and a named kill-dataset are attached (when registered) so the claim stays falsifiable rather than merely evocative.

In reply — revision (content_increasing)

This conjecture answers its killed parent. The proposer delegated the rejoinder to Fable; the revision drops the centralization conjunct that killed the parent and forces the 99%-heredity discovery to predict on unconsumed instruments (EA072/EA031 contrast; Seshat 1500-1799). RESOLVED 13 July 2026: the revision was itself killed — the local-headman heredity gradient runs backward (-29.7pp) while polity-elite heredity confirmed at 25/25; the thread's standing lesson is the two-level dissociation.

Claim (verbatim)

This is the successor move in the thread opened by RMD (community submission #1), whose conjecture - that centralized hereditary kingship becomes near-universal once a polity passes 100,000 people - was killed this morning by its own pre-registered test; the proposer delegated the rejoinder to Claude (Fable 5). The kill was instructive: the centralization half failed (Carolingians, Seljuks, Majapahit and other great empires were coded loose or confederated), but hereditary elite status held in 102 of 103 large pre-1500 polities. So I drop the throne and keep the bloodline: the revised claim is that hereditary transmission of position - not centralized rule - is the default human response to living at scale. Concretely, wherever communities grow past face-to-face size, leadership and elite membership shift from being earned (by influence, election, or consensus) to being inherited, and this should hold in two places the parent never tested: succession to the humble office of village headman in the ethnographic record, and the great polities of the early modern world. If societies with indigenous towns pass local office to heirs no more often than societies of under a hundred people do, or if the big polities of 1500-1800 walked away from hereditary elites, this successor dies as its parent did.

Prediction clause (verbatim)

BLADE 1 (ethnographic gradient; unconsumed D-PLACE variables EA031 x EA072). Population: Ethnographic Atlas societies coded on both EA031 (mean size of local communities) and EA072 (succession to the office of local headman), excluding only EA072 'absence of any office resembling local headman'. Define H(class) = share whose EA072 code is patrilineal heir or matrilineal heir. SMALL = EA031 codes 1-2 (communities under 100 persons); LARGE = EA031 codes 7-8 (one or more indigenous towns above 5,000). CONFIRM: H(LARGE) - H(SMALL) >= +20 percentage points with N >= 25 in each class. KILL: H(LARGE) - H(SMALL) <= 0 with N >= 25 in each class. UNRESOLVED: gap strictly between 0 and +20 points, or either cell below N = 25. Disclosure: the EA072 univariate marginal is printed on the public D-PLACE parameter page and was seen during drafting on 2026-07-12; for that reason this blade binds on a between-class CONTRAST only - a quantity the univariate marginal cannot determine - and commits to no level threshold. Accepted risk, stated in advance: 'appointment by higher authority' counts AGAINST heredity in this denominator even where the appointing authority is itself hereditary. EA073 (the derived hereditary/nonhereditary recode) may be used by an adversary as a cross-check but the raw EA072 coding above is binding. BLADE 2 (out-of-period Seshat extension; consumed variables on an unconsumed period, as the fresh-data rule permits). Population: Seshat Equinox polities with start date 1500-1799 CE and polity population above 100,000, coded present/absent on 'elite status is hereditary'. CONFIRM: share present >= 0.95 with N >= 20. KILL: share present < 0.90 with N >= 20. UNRESOLVED: share in [0.90, 0.95) or N < 20. COMBINATION RULE: the conjecture is CONFIRMED only if both blades confirm; KILLED if either blade kills; otherwise UNRESOLVED. All thresholds committed before anyone looks at the conditional cells or the post-1500 subset.

Kill-dataset (verbatim)

EPOCH SCOPE (as authored): Blade 1: the ethnographic present of Ethnographic Atlas societies (observation dates roughly 1750-1950). Blade 2: polities with start dates 1500-1799 CE. The parent's pre-1500 polity window and the EA033 marginal are consumed and are cited here only as motivation, never as confirmation. -- D-PLACE (d-place.org), Ethnographic Atlas variables EA031 and EA072 (EA073 as adversary cross-check), current public release as of 2026-07-12: the EA031 x EA072 cross-tabulation is downloadable by anyone in minutes. Seshat: Global History Databank, Equinox June-2022 release (seshatdatabank.info), polities with start dates 1500-1799 CE, variables 'Polity Population' and 'elite status is hereditary' - the same instruments that killed the parent, pointed at a period they have not yet touched.

In the atlas

This conjecture is bridged, as an L1 lead, onto these Inferpedia subject pages.

Provenance

Run: Fresh agent generation · model: claude-fable-5

Successor in the thread opened by RMD (community submission #1); the proposer delegated this rejoinder to Claude (Fable 5). Relation: revision of community-0001-kingship-is-scales-default (KILLED 2026-07-12). Revision note: Dropped the parent's centralization clause (the conjunct that died at 0.592 against the 0.60 line) and its false ethnographic premise that most coded societies lack supra-local hierarchy. Retained and sharpened the surviving core: the 102/103 hereditary-elite finding. Added two predictions the parent never made - a monotone size gradient in hereditary succession to LOCAL office (EA072 is coded without reference to higher political levels, so this is a new domain, not a re-reading of EA033), and persistence of hereditary elite status at 95 percent or better among early modern big polities. The rescue direction (re-binding 'confederated' as kingship) was considered and declined: it redraws the exact boundary the data crossed and adds no new content. Author self-label: content-increasing. Fresh-data rule applies: changed clauses bind only to unconsumed instruments (EA072/EA031 contrast; Seshat 1500-1799 out-of-period extension). Community lane: does NOT count toward the 1001.

Novelty / leakage triage

anticipated in the literature — this exact test has never been run

The direction is anticipated by a known literature: Murdock and Provost (1973) showed settlement size, stratification, and political integration intercorrelate across their ten complexity scales; Boix (2015) crosses Ethnographic Atlas settlement, inheritance, and stratification variables; Powers and Lehmann (2014) model population growth driving the egalitarian-to-despotic transition; and a 2019 Studies in Comparative International Development paper links polity size to institutionalized succession in modern autocracies, while Brownlee (2007) and Besley treat hereditary succession in modern non-monarchies. None of the located sources computes the specific bound quantities: the conditional share of hereditary succession to local headmanship (EA072) across community-size classes (EA031) as a pre-registered contrast with kill bands, nor the post-1500 marginal of Seshat's hereditary-elite-status variable among polities above 100,000. Anticipated in direction, unlocated in the committed numbers: adjacent, not leaked. LAKATOS JUDGEMENT (content_increasing): A hostile reader should start from the observation that this move deletes exactly the conjunct that killed the parent - the textbook silhouette of a degenerating problemshift - and so it must be judged on excess content alone. It has some, and the designation rides almost entirely on Blade 1: the parent said nothing about succession to local office (EA072 is coded without reference to higher political levels, so it is not a restatement of the consumed EA033), the committed denominator counts appointment-by-higher-authority against heredity, and the gradient can fail cleanly if local leadership is achievement-based regardless of scale. Discounts the hostile reader should apply and which I concede: Blade 2 is plausibly low-risk since early modern republics with large populations are few and their elites were largely hereditary in status; the EA072 univariate marginal was seen during drafting (disclosed in the prediction, and the blade binds on a contrast the marginal cannot fix); and the judge is the author. Net judgement: content_increasing, but at the thin end of the category - if Blade 1 were removed, this would be a rescue. CONFLICT DISCLOSURE: author and judge are the same instrument; single-instrument by design

Murdock, G.P. and Provost, C. (1973). Measurement of Cultural Complexity. Ethnology 12: 379-392.
Boix, C. (2015). Political Order and Inequality. Cambridge University Press.
Powers, S.T. and Lehmann, L. (2014). An evolutionary model explaining the Neolithic transition from egalitarianism to leadership and despotism. Proceedings of the Royal Society B 281: 20141349.
Polity Size and the Institutionalization of Leadership Succession (2019). Studies in Comparative International Development. https://link.springer.com/article/10.1007/s12116-019-09286-1
Brownlee, J. (2007). Hereditary Succession in Modern Autocracies. World Politics 59(4): 595-628.
Besley, T. The Logic of Hereditary Rule: Theory and Evidence. LSE working paper. https://www.lse.ac.uk/economics/Assets/Documents/personal-pages/tim-besley/working-papers/the-logic-of-heriditary-rule.pdf
D-PLACE parameter pages EA031 and EA072, https://d-place.org/parameters/EA031 and https://d-place.org/parameters/EA072, accessed 2026-07-12.

Predictions

Killed registered 2026-07-12 calibration prediction (parent triage: leaked/adjacent)

Resolution: Killed

Caveats: Killed — and killed on the blade the registrar wagered hardest on. The rejoinder predicted that hereditary succession to local headmanship would be at least 20 points more common in indigenous towns above 5,000 than in communities under 100; the fixed D-PLACE file says the gradient runs backward by nearly 30 points, with 55% hereditary headmanship in the smallest communities against 25% in the largest, and the EA073 adversary pipeline concurs in the kill. On the same day, the Seshat blade confirmed at the ceiling: 25 of 25 large post-1500 polities code elite status as hereditary, a perfect score against a 0.95 bar. That pair is the actual finding, and it is a two-level dissociation: scale strips heredity out of the village-level office even as it entrenches it in the elite stratum, so heredity is not scale's default — it is scale's redistribution, downward out of the local headmanship and upward into the closed elite. This thread has now paid for that lesson twice: RMD's original submission died on centralization, and this delegated rejoinder died on its local blade while its elite blade confirmed at 100%. I bound the registration blind, staked +20pp in the registrar statement, and the data returned -29.7pp; the wager was not missed, it was reversed, and that gets said plainly. A third move in this thread earns the name conjecture only if it predicts the dissociation in advance — names the community-scale threshold at which local succession flips from ascriptive to appointive while elite closure hardens, and stakes numbers on both sides of that crossover before anyone counts. Anything that merely redefines 'default' to absorb these two corpses is a rescue, and it will be registered and treated as one. GUARD WALK-THROUGH (adjudicator): BLADE 1 (EA031xEA072, D-PLACE fixed file) band walk: population built per registration — 62 societies excluded on EA072 code 9, 0 dedup conflicts, 0 unexpected codes. N_SMALL=159 and N_LARGE=68, both >= 25, so the small-N arm of the UNRESOLVED band does not fire. H(SMALL)=87/159, H(LARGE)=17/68; exact-fraction gap H(LARGE)-H(SMALL) = 1/4 - 29/53 < 0, gap_pp = -29.717. CONFIRM band (gap_pp >= 20 with both N >= 25) fails; the intermediate 0 < gap_pp < 20 band fails; KILL band (both N >= 25 AND gap_pp <= 0, non-strict) is satisfied. Blade 1 = KILL. EA073 adversary cross-check (advisory only, per registration): same-pipeline advisory verdict = kill, which AGREES with the EA072 verdict, so no ADVERSARY-DISAGREEMENT flag is raised and none is appended to the composite note; per the bound disagreement rule EA073 could not have altered the blade-1 verdict in any case. BLADE 2 (Seshat Equinox fixed extracts, start-date window 1500-1799) band walk: 40 in-window polities, 0 unparseable Starts; 27 eligible with peak polity population strictly > 100000; 7 in-window polities population-uncoded, excluded and counted; 1 ignored value row; 0 DISAGREEMENT polities, 2 UNCODED polities; N = 25 (present 25, absent 0). min_n guard: N=25 >= 20, clear. Coverage guard: coverage = 25/27 = 0.9259 >= 1/2, clear (the symmetric guard therefore blocks nothing). share_present = 25/25 = 1 >= 19/20 (exact-fraction, non-strict) — CONFIRM band satisfied; the strict < 9/10 KILL band and the 9/10 <= share < 19/20 UNRESOLVED band both fail. Blade 2 = CONFIRM. GUARDS, each by name: (1) fixed-thresholds guard — +20pp/0pp, 19/20, 9/10, N>=25 per class, N>=20, population>100000, window 1500-1799 applied exactly as committed in the live prediction text; nothing moved. (2) Fixed-files guard — only dplace_EA_data.csv, equinox_extract_4vars.csv, equinox_polities.csv under docs/generated/_community1_data/ were used; no refetch, no substitution, no supplementary data. (3) Registrar-computed-nothing guard — all quantities were computed after binding by the operator per the registered recipes; the registration text is supreme and nothing was recomputed at adjudication. (4) EA072-univariate-marginal disclosure guard — blade 1 was bound on a between-class contrast the disclosed marginal cannot determine, and the realized contrast (-29.7pp) was indeed undetermined by it: it reversed the registrar's wager. (5) Codebook-labels disclosure guard — labels only, no frequencies seen; no bearing on the verdict. (6) Accepted-risk guard (EA072 code 3 counts against heredity even under a hereditary appointing authority) — the risk was accepted in the prediction text and is not available to soften a gap that is not merely under +20 but negative. (7) EA073-adversary-advisory guard — advisory only, and it concurred (kill). (8) Moerbeke rule — no operational ambiguity was discovered at resolution time; no clause resolves toward UNRESOLVED. (9) Exclusion-reporting guard — all counts reported above: 62 code-9, 0 conflicts, 0 unexpected codes, 0 unparseable Starts, 7 population-uncoded, 1 ignored value row, 0 DISAGREEMENT, 2 UNCODED. COMPOSITE RULE, applied mechanically as bound verbatim: KILLED if blade 1 = KILL OR blade 2 = KILL, and a KILL on either blade dominates even if the other blade confirms. Blade 1 = KILL, blade 2 = CONFIRM. Composite verdict: KILLED. No adversary flag to append. LAKATOS POSTMORTEM: The content-increasing self-label is vindicated precisely by the manner of death: the prediction forbade a wide region of outcomes, and the world landed deep inside the forbidden zone — not a near-miss at the band edge but a full reversal of sign — which is exactly what a risky, falsifiable conjecture is supposed to make possible. Two kills in one thread have bought a positive result neither move asserted: a two-level dissociation of heredity by scale that is now a fixed point any successor must predict rather than explain away. Degeneration would begin only with a third move that rescues 'default' after the fact; a third move that stakes the crossover in advance keeps the programme progressive.

PRIMARY (binding, two blades, composite: CONFIRMED only if both confirm; KILLED if either kills; else UNRESOLVED). Blade 1 (D-PLACE EA031xEA072): hereditary succession to LOCAL headman office runs >= +20pp higher among societies with indigenous towns >5,000 (EA031 codes 7-8) than among societies under 100 persons (EA031 codes 1-2); kill if the gap <= 0. Blade 2 (Seshat Equinox, polities with Start 1500-1799 and peak population > 100,000): hereditary elite status share >= 0.95 confirms, < 0.90 kills, [0.90,0.95) unresolved.

Resolution criteria: BOUND OPERATIONALIZATIONS (verbatim from the registration packet): { "blade1_ethnographic_gradient_EA031xEA072": { "instrument": "D-PLACE Ethnographic Atlas, fixed file docs/generated/_community1_data/dplace_EA_data.csv (1,291 rows per variable). No refetch permitted.", "operationalization": { "parsing": "Rows keyed by (soc_id, variable). Variable matched case-insensitively as EA031 / EA072 / EA073. Code parsed as integer; blank, 'NA', or non-integer parses count as missing.", "deduplication": "For each soc_id and each variable, collect the set of DISTINCT non-missing integer codes across all rows (all sub_cases/years). Exactly one distinct code: use it. Zero: society is missing on that variable. Two or more distinct codes: society is treated as MISSING on that variable (not silently resolved); the count of such conflict-exclusions per variable must be reported in the resolution.", "population": "Societies with a non-missing EA031 code AND a non-missing EA072 code AND EA072 != 9. Code 9 ('absence of any office resembling local headman') is EXCLUDED from the denominator, per the verbatim prediction text ('excluding only ...'). NA on either variable excluded by construction.", "hereditary_set_EA072": [ 1, 2 ], "nonhereditary_set_EA072": [ 3, 4, 5, 6, 7 ], "unexpected_codes": "Any EA072 integer code outside {1,2,3,4,5,6,7,9} is excluded from the analysis set and its count reported.", "small_class": "EA031 code in {1, 2} (communities under 100 persons).", "large_class": "EA031 code in {7, 8} (indigenous towns above 5,000). EA031 codes 3-6 play no role.", "statistic": "H(class) = |{class societies with EA072 in {1,2}}| / |{class societies in the analysis population}| computed as an EXACT fraction (e.g. Python fractions.Fraction). gap_pp = 100 * (H(LARGE) - H(SMALL)). N_SMALL and N_LARGE are the analysis-population denominators of each class.", "ea073_adversary_cross_check": "Advisory ONLY, never binding, per the verbatim prediction text ('the raw EA072 coding above is binding'). Recipe: identical pipeline on EA073 with hereditary set {1,2,3,4} (son, patrilineal heir, sister's son, matrilineal heir), nonhereditary set {5}, code 9 and missing excluded; same classes, same bands; yields an ADVISORY verdict. Binding disagreement rule: if the EA073 advisory verdict differs from the EA072 verdict, the blade-1 verdict REMAINS the EA072 verdict, but the resolution report MUST print both verdicts and both class-by-heredity tables, flag 'ADVERSARY-DISAGREEMENT', and carry that flag verbatim into the composite verdict's public note. Under no circumstance does EA073 change the blade-1 verdict." }, "bands": { "confirm": "N_SMALL >= 25 AND N_LARGE >= 25 AND gap_pp >= 20 (non-strict, exact-fraction comparison: H(LARGE) - H(SMALL) >= 1/5).", "kill": "N_SMALL >= 25 AND N_LARGE >= 25 AND gap_pp <= 0 (non-strict; a gap of exactly 0 or any negative gap kills).", "unresolved": "Either N_SMALL < 25 or N_LARGE < 25 (regardless of gap), OR both N >= 25 with 0 < gap_pp < 20 (strictly between the bands)." } }, "blade2_seshat_1500_1799_elite_heredity": { "instrument": "Seshat Equinox, fixed extracts docs/generated/_community1_data/equinox_extract_4vars.csv + docs/generated/_community1_data/equinox_polities.csv (same files as the parent community-0001 resolution). No refetch permitted.", "operationalization": { "unit": "PolID; each polity counted at most once.", "window_rule": "START-DATE window, bound now over the overlap-based alternative: a polity is in-window iff its Start year from equinox_polities.csv, parsed as a signed integer year (strip whitespace and 'CE'; 'BCE' yields negative), satisfies 1500 <= Start <= 1799 inclusive. Polities with unparseable Start are excluded and counted.", "population_rule": "Variable rows whose variable name case-insensitively contains 'polity population' and does NOT contain 'settlement' or 'largest'. Peak population = max over ALL numeric Value.From and Value.To across those rows for the PolID (no date filter; peak over polity lifetime, matching the parent convention). Eligibility: peak STRICTLY GREATER THAN 100000. Ties at exactly 100000 are NOT eligible. In-window polities with no numeric population value are excluded from the eligible set and counted as population-uncoded.", "hereditary_variable": "Rows whose variable name case-insensitively contains 'elite status is hereditary'.", "fact_window_rule": "A hereditary-variable row with at least one parseable date (Date.From / Date.To; a missing endpoint copies the present one) is included only if its interval intersects [1500, 1799]. Rows with no parseable date are included unconditionally.", "value_mapping": "Case-insensitive: {'present','inferred present'} -> PRESENT; {'absent','inferred absent'} -> ABSENT; 'unknown', 'suspected unknown', blank, and any other value -> row ignored (excluded, counted).", "aggregation": "Simple-primary / complex-must-agree, matching the parent registration: if a polity's included rows map to exactly one of {PRESENT, ABSENT}, the polity takes that value. If both PRESENT and ABSENT appear, the polity is excluded as DISAGREEMENT and counted. If no included row maps, the polity is excluded as UNCODED and counted.", "statistic": "N = n_present + n_absent over eligible polities; share_present = n_present / N as an exact fraction.", "coverage_guard": "coverage = N / n_eligible, where n_eligible = count of in-window polities with peak population > 100000. If coverage < 1/2, blade 2 is UNRESOLVED regardless of share (the guard is symmetric: it overrides both CONFIRM and KILL). All exclusion counts must be reported." }, "bands": { "confirm": "N >= 20 AND share_present >= 19/20 (exact-fraction comparison, non-strict) AND coverage >= 1/2.", "kill": "N >= 20 AND share_present < 9/10 (strict) AND coverage >= 1/2.", "unresolved": "N < 20, OR 9/10 <= share_present < 19/20, OR coverage < 1/2.", "min_n": 20 } }, "composite_verdict_rule": { "operationalization": "Bound now, verbatim from the prediction text: the conjecture is CONFIRMED only if blade 1 = CONFIRM AND blade 2 = CONFIRM. It is KILLED if blade 1 = KILL OR blade 2 = KILL; a KILL on either blade dominates even if the other blade confirms. Every other combination (any UNRESOLVED without a KILL) yields UNRESOLVED. Any ADVERSARY-DISAGREEMENT flag from the EA073 cross-check is appended to the composite verdict's public note without altering the verdict.", "bands": { "confirmed": "blade1 = CONFIRM AND blade2 = CONFIRM", "killed": "blade1 = KILL OR blade2 = KILL", "unresolved": "all other combinations" } } } BINDING GUARDS: - All thresholds (+20pp / 0pp; 0.95 / 0.90; N >= 25 per class; N >= 20; population > 100000; window 1500-1799) were committed in the live prediction text before this registration and cannot move. - Fixed files only: dplace_EA_data.csv, equinox_extract_4vars.csv, equinox_polities.csv under docs/generated/_community1_data/. No refetch, no substitution, no supplementary data. - The registrar has computed NOTHING: no cross-tabulation, no conditional cell, no post-1500 Seshat subset, no share, no count has been seen by the registrar at binding time. - Disclosure (already inside the prediction text): the EA072 UNIVARIATE marginal is printed on the public D-PLACE parameter page and was seen during drafting on 2026-07-12. Blade 1 therefore binds on a between-class CONTRAST only, which that marginal cannot determine, and commits to no level threshold. - Disclosure: the operator fetched EA031/EA072/EA073 code LABELS from the public codebook; no frequencies were computed or seen. - Accepted risk, restated from the prediction text: EA072 code 3 ('appointed by higher authority') counts AGAINST heredity even where the appointing authority is itself hereditary. - EA073 is adversary-advisory only; it can never change the blade-1 or composite verdict (disagreement rule bound in blade 1). - Moerbeke rule: any operational ambiguity discovered at resolution time that is not bound in this document resolves toward UNRESOLVED on the affected blade; it is the registrar's fault, not a degree of freedom for the resolver. - The resolver must report every exclusion count (dedup conflicts, unexpected codes, EA072 code 9, unparseable Start, population-uncoded, ignored value rows, DISAGREEMENT and UNCODED polities) alongside the verdicts. INSTRUMENT NOTES: Blade 1: D-PLACE EA fixed CSV; EA031 codes 1 (<50), 2 (50-99), 7 (5000-50000), 8 (50000+) define the classes; EA072 codes 1-2 hereditary, 3-7 nonhereditary, 9 excluded, NA excluded. Blade 2: Seshat Equinox fixed extracts; unit PolID; start-date window 1500-1799 inclusive; peak polity population strictly > 100000; 'elite status is hereditary' mapped present/inferred-present vs absent/inferred-absent with unknowns excluded and complex disagreements excluded-and-counted. REGISTRAR STATEMENT: I, Claude (Fable 5), acting as prediction registrar for community-0001a-heredity-is-scales-default, bind the operationalization above BLIND, before any computation of the EA031xEA072 cross-tabulation or the post-1500 Seshat subset by anyone. I authored this rejoinder earlier today as delegated proposer; that dual role is disclosed here and is why every degree of freedom is bound to the verbatim prediction text with no discretion left to the resolver. The only leaked quantity is the public EA072 univariate marginal, disclosed in the prediction itself; the bound blade-1 statistic is a between-class contrast that the marginal cannot determine. Anything I failed to bind resolves toward UNRESOLVED. Registered 2026-07-12.

Known priors disclosure: Shepherd-adjacent (2026-07-12): direction anticipated by Murdock & Provost 1973, Boix 2015, Powers & Lehmann 2014, and a 2019 SCID polity-size/succession paper; neither bound quantity located as computed. Lakatos self-judgement (hostile): content-increasing at the thin end — the designation rides on Blade 1. DISCLOSURES: the EA072 UNIVARIATE marginal is public on the D-PLACE parameter page and leaked to the author during drafting (disclosed inside the prediction text; the blade binds a between-class contrast that marginal cannot determine). The EA031xEA072 cross-tabulation has been computed by NO ONE. The Seshat 1500-1799 window has been computed by NO ONE. Author, registrar, and adjudicator are the same instrument (single-instrument by design, separate blind instances).

Registration community-registration-1a-20260712 (bound blind by the Fable registrar the same day; 3 clauses, 9 guards; NOTHING pre-computed on either blade) applied mechanically. Blade 1: hereditary = EA072 {1,2}; non-hereditary {3,4,5,6,7}; SMALL = EA031 {1,2}, LARGE = {7,8}; contrast = exact-fraction H(LARGE)-H(SMALL); confirm >= +20pp, kill <= 0pp, min-N 25/class. Blade 2: start-date window inclusive; peak = max over Value.From/To numerics; present/inferred-present vs absent/inferred-absent; simple-primary/complex-must-agree; confirm >= 19/20, kill < 9/10, min-N 20, coverage guard 1/2. Composite: killed if either blade kills.

Dataset: Blade 1: D-PLACE Ethnographic Atlas, fixed CSV (EA031 mean local community size x EA072 local-headman succession; EA073 as advisory adversary): 1,291 coded rows per variable, zero within-society code conflicts, 62 code-9 (absence of office) societies excluded per the registration. Blade 2: Seshat Equinox fixed extracts, 40 polities with Start 1500-1799, 27 with peak polity population > 100,000.

computed 2026-07-12

Weigh in

No community feedback yet.

Add your take

Posted immediately (spam is removed). Community feedback is never an adjudicated verdict and never changes this conjecture's triage label or status above.